THEORY: non-concatenative morphology [was: Re: Ke'kh]

On Thu, 21 Sep 2000, Mangiat wrote:

> What's a non-concatenative grammar? One which does not rely exclusively on
> pre/suffixes but changes the root itself (or, à la Sémitique, the vowels in
> the root)?
Exactly. "Non-concatenative" suggests that a complex word isn't made
up by simply stringing together a collection of morphemes. Non-
concatenative morphology is usually considered part of prosodic
morphology, which also includes infixation, reduplication, and other
goodies. These are considered "prosodic" since the syllable/foot
structure imposes requirements on the correct expression of
morphological categories. For instance, infixation is really
prefixation which seeks to avoid creating syllable codas; in languages
which indulge in infixation, "infixes" which are attached to
vowel-initial stems show up as prefixes:

        um + tawag -> tumawag   'call, pf. actor trigger'
        um + aral -> umaral     'teach'

(I should say here that the correct interpretation of the data is not
uncontroversial; is _aral_ really vowel-initial, or is there a glottal
stop there?)

Reduplication is also considered prosodic since it is usually the case
that the reduplicant fills an "authentic unit of prosody" i.e., light
syllable, heavy syllable, foot, etc.

Many of the Semitic patterns are also prosodic in just this way; the
root and vocalic patterns are mapped to "authentic units of prosody"
such as iamb (light syllable-heavy syllable) or moraic trochee (two
light syllables or one heavy syllable), often with stuff left over on
the right edge of a form. The Arabic broken plurals come to mind:

         sg        pl

        nafs     nufuus       'soul'
        ?asad    ?usuud       'lion'
        rajul    rijaal       'man'
        jundub   janaadib     'locust'

For these nouns, the plural form must have an iamb; the plural of
'locust' _janaadib_ has an iamb and then some more: (ja)(naa)dib.

Dirk

--
Dirk Elzinga
dirk.elzinga@m.cc.utah.edu